Operating on Tangle
Operators run Blueprint services and earn fees for reliable execution. This section covers how to deploy and maintain operator infrastructure.
Quick Start
- Set up the Blueprint Manager: Manager Setup
- Configure your keystore: Key Management
- Register on-chain: Join as Operator
- Monitor your services: Quality of Service
What Operators Do
- Run the Blueprint Manager to execute service instances.
- Submit heartbeats to prove liveness.
- Respond to job calls and submit results on-chain.
- Maintain uptime to avoid QoS degradation and potential slashing.
Choosing Isolation
Operators select how to isolate workloads:
| Method | Security | Performance | Use Case |
|---|---|---|---|
| Native (no sandbox) | Low | Highest | Trusted blueprints only |
| Cloud-hypervisor VM | High | Good | Production default |
| Kata Containers | High | Good | Kubernetes deployments |
See Requirements for setup details.
If you want to serve TEE-aware workloads, continue with Confidential Compute.
Earning and Risks
Operators earn from:
- Service fees: Paid by customers, split with developers and stakers.
- Optional TNT incentives: When governance funds incentive pools.
Operators risk slashing for:
- Missing heartbeats beyond the grace period.
- Submitting invalid or late results.
- Violating isolation guarantees.
Sandbox Runtime Hosting (Early Access)
Operators can also host the sandbox runtime that executes agent workloads. The runtime schedules by available capacity, optionally pools and autoscales hosts, and exports Prometheus metrics per host.
Learn more in Runtime Architecture.
Next Steps
- Operator Runbook: Daily operations checklist.
- Pricing: How to set competitive rates.
- Benchmarking: Test blueprint performance.